nut are solar radiation data?

Solar radiation data provide information

on how much of the sun's energy strikes a
surface at a location on earth during a par-
ticular time period. The data give values of
energy per unit of area. By showing natural-
ly occurring changes in the amount of solar
radiation over the course of days, months,
and years, these data determine the
amount of solar radiation for a location.
The units of measurement are expressed
as kilowatt-hours per square meter
(kWh / m 2), megajoules per square meter
(Mj/m'), langleys (L), or British thermal
units per square foot (Btu/ ft').
Today, the primary source of solar radia-
tion data for the United States comes from
measurements made by the National
Weather Service at 26 SOLMET (SOlar
METeorological) stations from 1952 to
1975. In addition, mathematical models
estimated data for 222 ERSATZ (synthetid
stations where no solar radiation measure-
ments were made. Because the equipment
did not always accurately measure the
solar radiation and the models used. were
limited in their application, the data do not
always correlate well with
more recent field measure-
ments. To provide better data,
we developed a National
Solar Radiation Data Base.
This data base covers 30 years
(1961-1990) and comes from
information recorded by
more accura te instruments
and from better models. In
1992, this new data base will
be available for 250 sites.

4
 Guantanamo Bay, Cuba
Koror Island. Pacilic
Kwajalein ISland. Pacific
San Juan. Puerto Rico
Wake Island. Pacilic

" SOlMET
• ERSATZ

From 1952 to 1975 solar radiation was measured

at 26 SOLMET stations ( .. ) and modeled for 222
ERSATZ stations ( • ). Most of these stations will
be included in the new National Solar Radiation
Data Base.

5
 do we need solar radiation data?
The earth receives a vast amount of energy
from the su n in the fonn o f solar radiation.
If we converted to usable energy just 0.2%
o f the solar radiation that falls on our na·
tion, we would meet the energy demand of
the entire United States. A variety of solar
energy technologies are being developed to
harness the sun's energy including:
• solar electric (photovoltaid for convert-
ing sunlight directly into electricity;
• solar heat (thennaD for heating water for
industrial and household uses;
• solar thermal electric for producing
steam to run turbines that generate
electricity;
• solar fuellechnologies for converting
biomass (plants, crops, and trees) into
fuels and by-products;
• passive solar for lighting and heating
buildings; and
• solar detoxification for destroying haz-
ardous waste with concentrated sunlight.
6
"The more accurately rue knolU the
solar resource, the better we call
optimize the system. Therefore,
accurate soffir radiatio1J data are
011important factor ill solar
system design."
o.wld F. u.nkuc:d
~ NationIII..abonIIOrte.
The economics of these technologies
depend on the equipment and operating
costs, the percentage of the soL.1r rad iation
that can be converted into the desired
energy product, and the amount of solar
radiation available. Users of these tech-
nologies need h igh-quality solar radiation
data. If the actual solar radiatio n for a loca-
tion is less than indicated by available data,
the perfonnance and the economic goals
for the system will not be met. On Iheother
hand, if the actual solar energy at a location
is greater tha n indicated by the data, the
perfonnance and economic projections
may be 100 conservative and prevent a
viable technology from being used.
To minimize energy consumption, heat-
ing and air conditioning engineers also
use solar rad iation data 10 select building
configurations, orientations, and air con-
ditioning syslems. Because energy costs
are a significant expense in building owner-
shi p, an energy-efficient design can s ig-
nificantly reduce the life-cycle cost of a
building.
The amount of solar radiation received
changes throughout the day and year
due to weather patterns and the changing
position of the sun. and solar radiation
data reflects this variabili ty. By knowing
the va riability, we can size storage systems
so they can provide energy at night and
during cloudy periods. For technologies
 ..
• .!.. " .
J_ )-'

- u_ '~

Because of absorption and scattering by the atmo-

sphere, the spectral distribution of solar radiation
outside the ahnosphere differs significant1y from
thai o n earth. Also, the spectral dislribut.i on on
earth changes throughout the day and year and
is influenced by location, climate, and atmospheric
conditions. Consequently, the percentage o f energy
that is composed of UV, visible, or neal'-infrared
radialion, or portions thereof, 3.150 varies by
location, time of day, and year.

wi th no energy storage, we can eva ll1ate

load matching by comparing the profile o f
the available solar radiation throughout the
day w ith the profil~ of the energy required
by the load. Solar radiation data also help
determine the best geographic loca tions for
solar energy technologies. Other factors
being equal, a site receiving mo re solar
radiation will be more economical.
For certain technologies, we also need to
Th is remote wate ....leve l-monitoring station uses photovoliaics for charging
know the spectral, or wavelength, distribu- storage batteries that supply electric power. Solar radiation data provide
tion of the solar radiation. For example, information for determining the size of the photovoltaic and battery system
photovol taic devices respond primarily to needed to supply remote stations like this with reliable elecb'ic service.
wavelengths in the visible and near-in-
frared region of the spectrum, w hile solar
d etoxifi cation uses energy from the
ultraviolet (UV) region. Location, climate,
and atmospheric conditions influence the
spectral distribution of solar radia tion.

7
nu~ influences the amount of solar radiation?
The amount of solar radiation reaching the
earth's surface varies greatly because of
changing atmospheric conditions and the
changing position of the sun, both during
the day and throughout the year. Clouds
are the predominant atmospheric condition
that detennines the amount of solar radia-
tion that reaches the earth. Consequently,
regions of the nation with cloudy climates
receive less solar radiation than the cloud-
free desert climates of the southwestern
United States. For any given location, the
solar radiation reaching the earth's surface
decreases with increasing cloud cover.
Local geographical features, such as
mountains, oceans, and large lakes, in-
fluence the formation of douds; therefore,
the amount of solar radiation received for
these areas may be different from that
received by adjacent land areas. For ex-
ample, m01.Ultains may receive less solar
radiation than adjacent foothills and plains
located a short distance away. Winds blow-
ing against mountains force some of the air
to rise, and douds fonn from the moisture
in the air as it cools. Coastlines may also
receive a different amount of solar radia-
tion than areas further inland. Where the
changes in geography are less pronounced,
such as in the Great Plains, the amount of
solar radiation varies less.

8
 Many atmospheric scientists
believe that the eruption of
MOllnt Pinatubo in June 1991
will have worldwide effects
during the next few years.
This was one of the largest
volcanic eruptions of the 20th
century. Volcanic ash and
sulfur dioxide spewed high
above the Philippines and
into the stratosphere; the
resulting dust cloud spread
around the earth's equator
and toward higher latitudes.
The increased dust diminishes
the solar radiation received
at the earth's surface. Peak
effects will occur in 1992,
but colder winters and
cooler summers may ensue
until near the middle of
this decade. Long-term
measurement of solar
radiation at numerous sites
permits naturally occurring
events such as this to be
evaluated with respect to
their impact on the solar
resource and the climate.

Clouds are the predominant atmospheric condition that detennines the amount
of solar radiation reaching the earth.
The amount of solar radiation also varies
depending on the time of day and the
season. In general, more solar radiation is
present during midday than during either
the early morning or late afternoon. At mid-
day, the sun is positioned high in the sky
and the path of the sun's rays through the
earth's atmosphere is shortened. Conse-
quently, less solar radiation is scattered or
absorbed, and more solar radiation reaches
the earth's surface. In the northern hemi-
sphere, south-facing collectors also receive
more solar radiation during m idday be-
cause the sun's rays are nearly perpen-
dicular to the collector surface. Tracking
collectors can increase the amount of solar
radiation received by tracking the sun and
keeping its rays perpendicular to the collec-
tor throughout the day. In the northern
hemisphere, we also expect more solar
radiation during the summer than during
the winter because there are more daylight
hours. This is more pronounced at higher
latitudes.
Both man-made and naturally occurring
events can limit the amount of solar radia-
tion at the earth's surface. Urban air pollu-
tion, smoke from forest fires, and airborne
ash resulting from volcanic activity reduce
the solar resource by increasing the scatter-
ing and absorption of solar radiation. This
has a larger impact on radiation coming in
a direct line from the sun (direct beam)
than on the total (global) solar radiation.
Some of the direct beam radiation is scat-
tered toward earth and is called diffuse (sky)
radia tion (global = direct + diffuse):Conse-
quently, concentrators that use only direct
beam solar radiation are more adversely
affected than collectors that use global solar
radiation. On a day with severely polluted
air (smog alert), the direct beam solar radia-
tion can be reduced by 40% whereas the
global solar radiation is reduced by 15% to

10
25%. A large volcanic eruption may
decrease, over a large portion of the earth,
the direct beam solar radiation by 20% and
the global solar radia tion by nearly 10% for
6 months to 2 years. As the volcanic ash
falls out of the atmosphere, the effect is
diminished, but complete removal of the
ash may take several years.

11
 parts of solar radiation are measured?
The total or global solar radiation striking a
collector has two components: (1) direct
beam radiation, and (2) diffuse radiation.
Additionally, radiation reflected by the
surface in front of a collector contributes to
the solar radiation received. But unless the
collector is tilted at a steep angle from the
horizontal and the ground is highly reflec-
tive (e.g., snow), this contribution is small.
As the name implies, direct beam radia-
tion comes in a direct line from the sun.
For sunny days with clear skies, most of
the solar radiation is direct beam radiation.
On overcast days, the sun is ob~ by
the clouds and the direct beam radiation
is zero.
Diffuse radiation is scattered out of the
direct beam by molecules, aerosols, and
douds. Because it comes from all regions
of the sky, it is also referred to as sky radia-
tion. The portion of total solar radiation
that is diffuse is about 10% to 20% for clear
skies and up to 100% for cloudy skies.

Some of the solar radiation entering the earth's abnosphere is absorbed and scattered. Direct beam
radiation comes in a direct line from the sun. Diffuse radiation is scattered out of the direct beam by
molecules, aerosols, and clouds. The sum of the direct beam, diffuse, and ground-reflected radiation
arriving at the surface is caUed total or global solar radiation.

12
 The type of data needed and the funds
available help determine the number and
kinds of instruments used at a site to
measure solar radiation. A complete solar
radiation monitoring station has instru-
mentation for meas uring three quantities:
(1) to tal or global radiation on a horizontal
s urface, (2) diffuse radiation on a horizontal
s urface, and (3) direct beam radiation.
Measuring all three quantities provides
sufficient information for understanding
the solar resource and for rigorous quality
assessment of the data . Any two of the
measured quantities can be used to calcu-
late a range of acceptable values for the
third . Many monitoring stations also have
equipment for measuring solar radiation on
tilted and tracking surfaces and for measur-
ing meteorological parameters such as am-
bient temperature, relative humidity, and
wind speed and direction.
A station with a lower level of funding
may only measure two quantities; the third
is calculated. For example, the direct beam
component can be derived by s ubtracting
the diffuse radiation from the global
radiation and applying trigonometTic
relationships to account for the position
of the sun. The trade-off for this approach is
that the calculated direct beam data are less
accurate than if the direct beam data were
measured . .
HistOrically, many stations have
measured only the global radiation on a
horizontal surface. This necessitates cal-
culating both the diffuse and direct beam
solar radiation, which results in less
accurate values for these two quantities
than if they were measured.
In the absence of any solar radiation
measurements, we employ models using
meteorological data such as cloudiness
and minutes of sunshine to estimate solar
radiation. Although much less accurate,
this is often the only option we have for
locations where solar radiation is not
measured. Cloudiness data are based on
observations by a trained meteorologist
who looks at the sky and estimates the
amount of cloud cover in tenths. A dear
sky rates a cloud cover value of 0 tenths,
and an overcast sky rates a cloud cover
value of 10 tenths. Minutes of sunshine are
recorded by an instrument that measures
the time during the day when the sun is
not obscured by clouds.
To investigate the spectral distribution
of solar radiation, an instrument called a
spectroradiometer measures the solar
radiation intensity at discrete wavelengths.
Spectroradiometers are complex and
relatively expensive instruments, and
their operation and maintenance require
Significant effort. Consequently, spectro-
radiometers are not routinely used for
long-tenn data collection. Rather, they
help establish data bases that have suffi-
cient information to validate models that
predict the spectral distribution based
on meteorological data and the position
of the sun.
13
 ow do we use solar radiation data?
Solar energy technologies rely on solar
radiation to provide energy for producing
electricity, heating water, destroying toxic
wastes, and lighting and heating buildings.
Common to these technologies is that the
end-use product is, for the most part, a
direct function of the amount o f solar radia-
tion received and the conversion efficiency.
Windows can significantly affect the heating and cooling loads of buildings.
Engineers and architects can use solar radiation data 10 evaluate the effects
that windows will have on the energy consumption of a building and hence
delennine the size of heating and air conditioning equipment needed.
16
That is, if the amount of solar radiation
is increased, then the end-use product
increases also. This is also true for solar
fuel production, in which crops are grown
and then converted into fuels and by-
products. Although d ependent on the
soil type and rainfa ll, crops also depend on
the amount of solar radiation received.
To determine the performance and eco-
nomics of sola r conversion technologies,
designers and eng ineers u se solar radiation
data to estimate how much solar energy is
available for a site. Depending on the par-
ticular technology, the solar collector might
be a photovoltaic array, a concentrating
parabolic trough, a domestic hot water
collector, a w indow, a skylight, or a canopy
of foliage. Desig ners and engineers use
hand calculations or computer simulations
to estimate the solar radiation striking a
collector.
Hand calculations are appropriate w hen
using solar radiation data that represent
an average for an extended period. For
example, designers of remote photovoitaic
powered systems for charging batteries use
average daily solar radiation for the month
to d etermine the size of the photovoltaic
array. The criteria for this application is not
the amount of solar radiation for a g iven
hour or day but w hether or not the average
daily solar radiation for the month is suffi-
cient to prevent the batteries from becom-
ing d ischarged over several days.
The month used in the design process
depends on the relative amount of solar
radia tion available compared to the energy
required by the load . For a system in w hich
the load is constant throughout the year,
solar rad iation data for December or
January are usually used for the northern
hemisphere.
 Computer simulations are an effective
tool when an hour-by-hour performance
analysis is needed. Utility engineers may
want to know if the output of a solar .
electric power plant could reliably and
economically help meet their daytime
electric demand. (One of the potential
benefits of a solar electric power plant is
that its output may coincide with the utility
peak electric demand for summertime air
conditioning loads.) By using the hourly
solar radiation data for its location, the
utility can run computer programs that
show how much energy could be
produced on an hour-by-hour basis o 4 8 12 16 20 24
throughout the year by the solar electric lime of day (hour)
power plant.
Computer simulation using solar radiation data shows how the output of two
Some solar energy conversion tech- photovoltaic power systems could be added to the utility's generation to help
nologies require a threshold value of solar meet peak electric demand in the summer. The fixed-tilt array faces south and
radiation before certain operations can is tilted from the horizontal at an angle equal 10 the site's latitude. The tracking
array uses motors and gear drives to point the array at the sun throughout the
begin or be sustained. As an example, a
day. Depending on location, the photovoltaic system with the 2-axis tracking
central receiver solar thermal electric array receives annually 25% to 40'7~ more global solar radiation than the fixed-
power plant may require direct normal tilt photovoltaic system and provides more power for longer periods. This
solar radiation values above 450 W 1m2 to must be weighed against the higher initial cost and maintenance required
for the tracker.
produce stearn for the turbine generator.
Consequently, to evaluate a site's potential
for solar thermal electric production, a
designer examines the solar radiation data
to determine the times of day when the "Because the solar load is the largest
solar radiation exceeds the threshold value. compO/len t for building exterior
Heating and air conditioning engineers
use solar tadiation'data to optimize build-
surfaces, and because windows are
ing designs for energy efficiency. For the most sensitive to the solar load,
example, window orientation and size can solar radiation data are essential for
affect the heating and cooling of the build-
the accurate and energtJ efficient
ing. South-facing windows transmit solar
energy in the winter that is beneficial in desigtl of buildings and their air
reducing heating requirements. But in the condition ing systems."
swnmer, solar energy transmitted through Jack F. Roberts, P.E.
windows (primarily those that face east or American Society of Heating, Refrigerating
and Air-Conditioning Engineers
west), must be offset by increased operation
of the air conditioning system. By having
access to solar radiation data for their loca-
tion, engineers and architects can evaluate
the effects of window orientation and size

17
 on the energy constunption of the building
and determine the size of the heating and
air conditioning equipment needed. They
can use this information, combined with
desired levels of natural lighting and the
building aesthetics, to fonnulate the final
building design.
Except for concentrator systems, solar
radiation data cannot be used without first
accounting for the orientation of the solar
collector. Concentrators track the sun and
focus only direct beam radiation, but flat-
plate collectors receive a combination of
direct beam radiation, diffuse (sky) radia-
tion, and radiation reflected from the ground
in front of the collector. Depending on the
direction the collector is facing and its tilt
from the horizontal, flat-plate collectors
receive different amounts of direct beam
radiation, diffuse radiation, and ground-
reflected radiation. Designers employ
equations to calculate the total or global
radiation on a flat-plate collector. The
equations use values of the direct beam
radiation, the diffuse radiation on a
horizontal surface, and the orientation
of the collector.
To maximize the amount of solar radia-
tion received during the year, flat-plate
collectors in the northern hemisphere face
south and tilt from the horizontal at an
angle approximately equal to the site's
latitude. The annual energy production
is not very sensitive to the tilt angle as long
as it is within plus or minus 15° of the
latitude. As a general rule, to optimize the
perfonnance in the winter, the collector can
Concentrator collectors (top) use direct beam solar radiation; flat-plate collectors be tilted 15° greater than the latitude. To
(bottom) use direct beam radiation, diffuse (sky) radiation, and ground-reflected optimize performance in the summer, the
radiation. collector can be tilted 15° less than the
latitude. Solar radiation data combined
with computer simulations can define
these relationships more precisely.
In the initial design stage, d esigners of
cells used in photovoltaic mod ules can
use spectral solar radiation da ta bases and

18
models to optimize the ceUs for maximum
energy production. Because the spectral
content of solar radiation changes through-
out the day and season, photovoltaic ceUs
are tailored for a specific range of solar
radiation wavelengths that will produce
the most energy. Different photovoltaic "For sizing stand-alone PV systems, we calclilate the
materials have different peak responses;
Ilwnber of PV modliles required to keep the batteries
performance models using spectral solar
radiation d ata bases can compare two or charged by lIsing the average daily solar radiation incident
mo re photovoltaic materials operating on the collector for the month of the year with the smallest
under a range of seasons and climates. ratio of solar radiation to electric load demand."
This results in optimizing the design early
RIchard N. Chapman
and eliminates the expense and time that SandIa NaUonal Laboratories
would o therwise be needed fo r prelimin-
ary field testing.

Stand-alone PV system
I Wor1<sheet It2 I Determine Design Current and Array Tilt worksheet fordetennining
the best collector tilt angle
I ~ ~I ~ I and the total charging
current required from

~,." r".., ~,~ ~~ ~ f"9,... f"_"

the PV modules. (As per
Stand-A fane PllOtovo ftaic
~ (t.~y) I ~~
(=J ~ (t.~ (....J I ~ Syste",s: A Handbook of
Recommended Desigtl
Practices,SAND87-1023,
II!:+: Albuquerque, NM: Sandia
National Laboratories,
March 1990.)

SeIed the largest Design Current and Corresponding Peak &.11rom each Lalilude and Enter Below

I~"."~
(hr~VJ I =
(aJT11$J
"" ~ ~
I1""1 (1Ys:.:1'l (ampSJ
"'" ~ ~
!>9
- ~JI (~J
4.44 10.2 5.21 8.6 5.74 7.9

Now Selecl the SmaUesl Oesigll Current aJ'ld Corresponding Peak Sun

Note: ~ ~ lIfI'ay is desired. use \racking data'rom Appendix A.

00 not mIX \l'ac:Img and lilced array data on !he $afTIII sheet £.J ~
"""",
5.74
....,
,~
7.9
SO'

19
ne1'e can you obtain solar radiation data?
The National Weather Service of the
National Oceanic and Atmospheric
Administration (NOAA) operates monitor-
ing stations in the United States to collect
and disseminate information about solar
radiation. This information is available on
compu ter readable magnetic tape from
NOAA's National Climatic Data Center
(NCOC), Federal Building, Asheville, NC
28801 (704) 259-0682.
Most of NOAA's solar radiation data
sets are from 26 SOLMET stations and
222 ERSATZ stations and consist of hourly
values of solar radiation and meteorologi-
cal data from 1952 to 1975. For theSOLMET
stations, instruments measured the global
horizontal solar radiation and researchers
modeled the direct beam solar radia tion
data. For the ERSATZ stations, although no
solar radiation measuremen ts were made,
researchers modeled global horizontal
radiation based on observed meteorologi-
cal data su ch as cloudiness and minutes of TMY data, researchers
sunshine; the ERSAlZ data do not include included values of direct
direct beam radiation. Because all the beam radiation with
ERSATZ data are modeled, these data are modeled values of global
less accurate than the SOLMET data. horizontal radiation. These data sets repre-
NOAA also has available more recent sent typical values occurring from 1952 to
data for the periods 1977 to 1980 and 1988 1975, and not the minimum or maximum
to the present. The data include hourly values. For example, a cloudy year in this
values of measured global horizontal solar period may have had an annual solar radia-
radiation for 38 s tations, measured direct tion value 10% below theTMYvalue, and a
beam solar radiation for 32 stations, and very cloudy month in this period may have
measured diffuse horizontal radiation for had a solar radiation value 40% percent
nine stations. below its TMY value. A difference between
Two of NOAA's data sets are of partic- TMY and WYEC data is that the TMY data
ular interest to designers and engineers: the are weighted toward solar radiation values
typical meteorological year (TMY) data set and their hourly distribution, whereas the
and the weather year for energy calculations WYEC data are weighted toward average
(WYEC) data set. For these, researchers monthly values of temperatures and solar
extracted infonnation from SOLMET/ radiation. Researchers recently revised
ERSATZ data to make data sets of hourly the WYEC data to include estimates of
values spanning one year. For the ERSATZ direct beam and diffuse solar radiation

20

and estimates of illuminance for lighting
applications. llluminance refers to solar
radiation in the visible region of the solar
spectrum to which the human eye
responds.
Solar radiation data derived from the
SOLMET/ERSATZ data sets are also pub-
lished in tabu lar fonn by the National
Technicallnfonnation Service (NTIS), U.s.
Department of Commerce, 5285 Port Royal
Road, Springfield, VA 22161. Two of these
tabular data sets are listed below.
Illsolatioll Data Manunl and Direct Nonnai
Sowr Rndiatioll Data Malmal, SERl / TP-22O-
3880, Gold en, CO: Solar Energy Research
Institute, July 1990.
This map shows the g loba l
solar radiation for a south·
facing collector tilled at an
angle equal 10 the site
latitude as an annual daily
average for different
locations in the United
States. one numbers on
the map represent MJ/m2;
multiply by 0.2778 to
obtain kWhlm1.)
This manual contains monthly averages
of global horizontal and direct beam solar
radiation, ambient temperature, the ratio
of global horizontal solar radiation on earth
to that outside the earth's ahnosphere (Kt),
and hea ting and cooling degree-days.
This info nnation is presented for all the
SOLMET/ERSATZ sta tions.
Stand-Alone PllOtovo/taic Systems: A
HandOOok of Recommended Design Prod ices,
SAND87-7023, Albuquerque, NM: Sandia
National Laboratories, March 1990.
The appendix of this handbook contains
monthly estimates of solar radiation strik-
ing collectors. These estimates are calcu-
lated for different tilts and sun-tracking
21

NOAA's National Oimatic Data Center has solar radiation data ava ilable on
computer readable magnetic tape. The data sets are for 26 SOLMET stations
and 222 ERSATZ stations and consist of ho urly values of solar radialion and
meteorological data from 1952 to 1975.
schemes. The estimates are for a selected set
of 38 SOLMET (ERSATZ stations and are
based on theSOLMET(ERSATZ data.
Maps are available that depict long-term
average solar radiation d ata for each month.
This is a convenient way to show variations
in the amount of solar radiation and for in-
terpola ting data between sta tions. For the
United States, these maps were made using
solar radiation data from the SOLMET /
ERSATZ data base. The Solar Rndiafioll
Energy Resource Atlas of tlte United States ItNlS
publis hed by the Superintendent of Docu-
ments, but is out of print. This a tlas is avail-
able a t some university and city libraries.
The University of Lowell compiled an
intemational solar radiation data base for
locations ou tside the United States. This
data base presents average d aily values
by month a nd year for global horizontal
solar radiation. It is ava ilable from the
University of Lowell Photovoltaic Program,
1 University Avenue, Lowell, MA 01854
(508) 934-3377.
22
Solar radiation data recorded for
l-minute intervals are available for four
locations: Albany, New York; Atlanta,
Georgia; Davis, California; and San Antonio,
Texas. The d ata were recorded over periods
of 1 year or more by university meteorologi-
cal research and training stations. Because
of the time scale used, these da ta are
primarily of interest to research ers study-
ing transient responses in solar energy tech-
nology systems. These data are available
from the National Renewable Energy
Labora tory (NREL), 1617 Cole Boulevard,
Golden, CO 80401.
A spectral solar radiation data base repre-
senting a range of atmospheric and climatic
conditions is also available fro m NREL.
This da ta base includes more than 3()(x)
spectra measured over a wavelength
range from 300 to 1100 nanometers at
2-nanometer increments (1 nanometer is
one-billionth of a meter) and is the result
of a cooperative effort between NREL,
the Electric Power Research Institute, the
Rorida Solar Energy Center, and the Pacific
Gas a nd Electric Company. Spectral solar
radiation was measured at three sites: Cape
Canaveral, R orida; San Ramon, California;
and Denver, Colorado. This data base can
help determine w hether spectraUy selective
technologies (such as photovoltaics and
biomass) are optimized for a particular
location and climate.
Additionally, other sources of solar radia-
tion d ata are state and local governments,
utilities, and universities. Examples include
the Padfic Gas and Electric Solar Insolation
Monitoring Program, the University of
Oregon/Pacific Northwest Solar Radiation
Data Network, and the Historically Black
Colleges and Universities Solar Radiation
Monitoring Network.
 ow accurate do the data need to be?
The required ac-
curacy of the solar
radiation data for
a site depends on
the application.
When the cost of the solar conversion
device is low compared with the overall
system cost, we can account for uncertain-
ties in the solar radiation data by using
"engineering judgment" to increase the
size of the solar collectors. However, as the
"Utility engineers need solar
radiation data accurate to within
±5% to assess the resource, estimate
the output of a solar system, and
determine whether the system can
reliably and economically meet
daytime demand and energtJ
requirements. Because there are few
sites with data of this accuracy, we
need monitoring stations to collect
the data at proposed PV sites."
J.E. Bigger
Electric Power Research Institute
accurate to within 5%. This is due to the
measurement uncertainties of the instru-
ments used and the limited number of
measurement sites. Consequently, desig-
ners today have to apply these data more
conservatively than is ultimately desirable.
The SOLMET/ERSATZ data are the
most widely used solar radiation data. On
an average for all sites, they are accurate to
within about 10% for average daily values
on an annual basis. But for average daily
values on a monthly basis for an individual
s ite, they can be in error by 20% or more.
For interpolating data for sites between
SOLMEf /ERSATZ stations, microclimate
differences due to terrain and local weather
conditions can also increase the uncertainty
of the data.

solar energy conversion system increases in

size and cost, this becomes less acceptable,
and we need more accurate solar radiation
data to optimize the design and project
the cost.
For large-scale applications of solar
energy conversion technologies, most For large-scale applications, like this 6.5-MW photovoltaic system, designers
experts agree that solar radiation data prefer solar radiation data that are accurate to within 5% so they can make sound
should be accurate to within 5% so they assumptions concerning system output, performance, and economics. (Photo
courtesy of Siemens Solar Industries')
can make reasonable assumptions con-
cerning energy output to evaluate the per-
formance and economics. Unforhmately,
not much available solar radiation data are

23
awwillwemeet our solar radiation data needs?
One of the goals of the Solar Radiation
Resource Assessment Project at NREL is to
provide accurate information about solar
radiation to minimize the economic risk
of implementing solar energy conversion
technologies. The data must accurately
represent the spatial (geographid, tem-
poral (hourly, daily, and seasonal), and
spectral (wavelength distribution) vari-
ability of the solar radiation resource at
different locations.
24
The new National Solar Radiation Data
Base (1961-1990) for the United States will
improve data quality over the existing
SOLMET /ERSATZ (1952-1975) data base.
For this new data base, NOAA used better
equipment for measuring solar radiation
at more sites and NREL used better model-
ing techniques for synthetic stations.
Scheduled for completion in 1992, this new
data base will include data for 250 sites.
After completing the data base we will
produce special purpose products such as
typical meteorological year (TMy) data
sets, maps, and data summaries.
By continuing the long-tenn measure-
ment of solar radiation at numerous sites,
we can assess changes in climate and add
new data to existing data bases. We can
improve the quality of the solar radiation
data base for the United States by working
with existing regional solar radiation
networks and establishing educational
initiatives so that data are being collected
at several hundred sites in the United
States. This large number of measurement
sites will improve the quality of the solar
radiation data base, better represent the
geographic distribution of solar radiation
in the United States, and provide research
data to develop techniques to estimate
solar radiation where there are no measure-
ment stations.
This type of research involves develop-
ing spatial interpolation techniques, such
as mapping so~ar radiation using cloud-
cover infonnation from satellites, to
estima te solar radiation between measure-
ment stations. This cloud-cover mapping
technique promises high spatial resolution
for the optimum siting of solar energy con-
version technologies and enables estimat-
ing solar radiation for countries where
no solar radiation data base exists.
 NREL is improving the equipment and
techniques used to measure solar radiation
and the models and methods used to deter-
mine the performance of solar conversion
technologies. Our recent activities include:
• angular response characterization and
uncertainty analysis of solar radiometers,
• development of improved quality assess-
ment procedures for solar radiation data,
• calibration of radiometers for industry
and members of the scientific community,
• development of both broadband and
spectral solar irradiance models, and
• conhibutions to the development of
solar trackers and spectroradiometers.
For information about solar radiation
data, models, and assessments contact Ooud-cover infonnation. analyzed from photographs
the NREL Technical Inquiry Service at taken by satellites, has the potential for estimating solar
303/231-7303. radiation at any location on earth.

25